Solar Under Storm Select Best Practices for Resilient Roof-Mount PV Systems with Hurricane Exposure - The Clinton Foundation
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
MOUN
KY
TA
ROC
IN IN
STIT UTE
Solar Under Storm
Part II
Select Best Practices for Resilient Roof-Mount PV Systems with Hurricane Exposure
BY CHRISTOPHER BURGESS, SANYA DETWEILER, CHRIS NEEDHAM, FRANK OUDHEUSDEN
AUTHORS & ACKNOWLEDGMENTS
AUTHORS ACKNOWLEDGMENTS
Christopher Burgess, Rocky Mountain Institute This report was made possible by The Clinton
Sanya Detweiler, Clinton Climate Initiative Climate Initiative’s funding from the Norwegian
Chris Needham, FCX Solar Agency for Development Cooperation, the Nationale
Frank Oudheusden, FCX Solar Postcode Loterij, and the players of the People’s
Postcode Lottery.
* Authors listed alphabetically
CONTRIBUTORS
Joe Cain, Solar Energy Industries Association
John Doty, UL
James Elsworth, National Renewable Energy Laboratory
Joseph Goodman, Rocky Mountain Institute (previously)
David Kaul, Salt Energy
Marc Lopata, Solar Island Energy
Dana Miller, ATEC Energy BVI
Fidel Neverson, Energy Solutions, Inc.
Edward Previdi, EP Energy
Carlos Quiñones, CJQ Engineering
Kevin Schnell, Caribbean Solar Company
Otto VanGeet, National Renewable Energy Laboratory
Angel Zayas, AZ Engineering
* Contributors listed alphabetically
CONTACTS
Christopher Burgess cburgess@rmi.org
Sanya Detweiler, sdetweiler@clintonfoundation.org
SUGGESTED CITATION
Burgess, C., Detweiler, S., Needham, C., Oudheusden,
F., Solar Under Storm Part II: Select Best Practices
for Resilient Roof-Mount PV Systems with Hurricane
Exposure, Clinton Foundation, FCX Solar, and Rocky
Mountain Institute, 2020. https://rmi.org/insight/solar-
under-storm/ and www.clintonfoundation.org/Solar-
Under-Storm.
Cover image courtesy of Sanya Detweiler, Clinton
Foundation
ABOUT US
MOUN
KY
TA
ROC
IN
IN
STIT UTE
ABOUT ROCKY MOUNTAIN INSTITUTE
Rocky Mountain Institute (RMI)—an independent nonprofit founded in 1982—transforms global energy use to
create a clean, prosperous, and secure low-carbon future. It engages businesses, communities, institutions, and
entrepreneurs to accelerate the adoption of market-based solutions that cost-effectively shift from fossil fuels to
efficiency and renewables. RMI has offices in Basalt and Boulder, Colorado; New York City; the San Francisco Bay
Area; Washington, D.C.; and Beijing.
ABOUT THE CLINTON FOUNDATION
The Clinton Foundation convenes businesses, governments, NGOs, and individuals to improve global health
and wellness, increase opportunity for girls and women, reduce childhood obesity, create economic opportunity
and growth, and help communities address the effects of climate change. The Clinton Climate Initiative (CCI)
collaborates with governments and partner organizations to increase the resilience of communities facing climate
change while reducing greenhouse gas emissions.
ABOUT FCX SOLAR
FCX Solar is an engineering consultancy and intellectual property development company focused on the PV
industry. It was founded in 2016 by Frank Oudheusden and Chris Needham, who together have a combined 25+
years in the PV industry. FCX Solar provides solar power developers and racking manufacturers with a wide range
of engineering services. FCX Solar has developed several products in the solar structure space and has a passion
for solving unique issues for its clients and partnerships.
TABLE OF CONTENTS 06 08 12 15 18 PREFACE................................................................................................................................................................ 06 EXECUTIVE SUMMARY.................................................................................................................................... 08 Summary of Findings.....................................................................................................................................................09 Recommendations........................................................................................................................................................09 1: INTRODUCTION................................................................................................................................................12 Approach....................................................................................................................................................................... 14 Organization.................................................................................................................................................................. 14 2: ROOT CAUSE IDENTIFICATION.................................................................................................................15 3: FAILURE MODE AND EFFECTS ANALYSIS ...........................................................................................18 The Additional Cost to Increase Resiliency................................................................................................................... 20
26 32 37 39 47 4: TECHNICAL DISCUSSION...........................................................................................................................26 Three-Rail Systems....................................................................................................................................................... 27 Joint Loosening / Top Down Clip Failure Analysis.......................................................................................................... 28 Modules Overhanging the Roof Edge........................................................................................................................... 29 Topographic “Speed-Up” Effects for Determination of Design Wind Speed................................................................... 31 5: CONCLUSION..................................................................................................................................................32 Recommendations........................................................................................................................................................ 34 Specifications................................................................................................................................................................ 34 Collaboration................................................................................................................................................................ 35 Energy Storage Systems for Resilience......................................................................................................................... 35 RECOMMENDED REFERENCES..................................................................................................................... 37 APPENDICES.........................................................................................................................................................39 APPENDIX A................................................................................................................................................................. 40 APPENDIX B.................................................................................................................................................................. 41 APPENDIX C................................................................................................................................................................. 44 APPENDIX D................................................................................................................................................................. 45 ENDNOTES............................................................................................................................................................ 47
PREFACE Solar Under Storm Part II is a response to the High wind speeds increase risk factors for solar overwhelming reception of the original report, which projects tremendously, but many solar installation provided best practices for ground-mount solar companies inadvertently overlook or incorrectly apply photovoltaic (PV) projects. It is also a response to low-wind speed designs (borrowed from Europe or stakeholder requests for a rooftop-focused report for the United States) for projects in high-wind zones the growing commercial and residential solar industry like the Caribbean. These low-wind mistakes become in the Caribbean and other vulnerable geographies catastrophic in high-wind events. with exposure to high-wind events. Image courtesy of Rocky Mountain Institute
PREFACE
Solar PV failure reporting is needed because some aimed at increasing the resilience of current and future
failures are highly visible while others are not, rooftop PV systems. This report will touch upon flat-
either because they are infrequent in occurrence or roof and pitched-roof PV power systems containing
because they are privately dealt with and not publicly flat-mounted, tilt-mounted, fully ballasted, and hybrid
published. Showcasing a wide range of failures has ballasted/penetrating systems. It excludes canopy PV
multiple benefits: systems and ground-mounted systems (both fixed and
tracking) as the recommendations for rooftop projects
• It provides proof to designers, installers, and are specific to their application. Canopy and tracking
customers that solar PV system resilience matters systems may be addressed in future versions of the
report if interest persists. Ground-mounted systems
• Ramifications for product and project design, vendor were addressed in the original Solar Under Storm
selection, installation, and maintenance become report, which is still available from Rocky Mountain
real because they are tangibly connected to real- Institute (RMI).1
world failures
This report is organized into five sections:
• It helps solar professionals learn from past mistakes, 1. Introduction
which is critical as repeating mistakes damages the 2. Root cause identification methodology and
reputation and credibility of the solar industry findings
3. Failure mode and effects analysis (FMEA)
4. Technical discussion
Like the first version, this report provides an opportunity 5. Conclusion
to address resilience for both a general and technical
audience. The report disseminates technical information The intended audience for Sections 2, 3, 4, and the
to non-technical readers and creates a more informed Appendix is engineering professionals responsible
solar professional, regulator, government official, utility, for PV system design, PV system specifications, and/
and customer. A well-informed customer base will or PV system construction oversight and approval.
systematically strengthen the PV industry by requiring Sections 1 and 5 are intended for a more general
vendors to incorporate resilience guidelines into their audience of customers, governments, utilities,
projects. In an industry that has experienced drastic regulators, developers, and PV system installers who
cost reductions year after year, in the “race-to-the- are interested in improving PV system survivability to
bottom” aspect of project and product design, it is intense wind-loading events.
critical for customers to understand best practices and
not accept low-cost shortcuts that could jeopardize Solar Under Storm Part II was developed with direct
project life or energy production. Supplying the feedback from solar companies in the Caribbean that
customer with a minimum set of guidelines raises the learned lessons in solar project resilience firsthand
bar, and those guidelines can only be improved through during and after Hurricanes Irma, Maria, and Dorian.
innovation and definitive testing, which in turn creates a Continuous feedback from the solar installer community
stronger industry. is vital to the success for solar PV resilience. Thus, RMI
and the Clinton Foundation’s Clinton Climate Initiative
The purpose of this document is to respond to the will host workshops and other opportunity for on-going
growing needs of the solar industry and combine field communication on this topic—notably through the forum
observations, photographic evidence, and expert of the Clinton Global Initiative (CGI) Action Network on
analysis to provide actionable recommendations Post-Disaster Recovery.2
SOLAR UNDER STORM PART II – 7
EXECUTIVE SUMMARY
Image courtesy of FortisTCI, Turks and Caicos
EXECUTIVE SUMMARY
The 2017 hurricane season was one of the most guide combines photographs from immediately after
active in history.3 Hurricanes Harvey, Irma, and Maria storms along with expert analysis to deliver actionable
caused widespread destruction throughout the recommendations for increasing resilience with rooftop
Caribbean. In 2019, Hurricane Dorian decimated the solar PV installations.
northern Bahamas bringing historic winds, rainfall, and
unprecedented destruction to the electricity system
and other critical infrastructure.4 In addition to the SUMMARY OF FINDINGS
emotional toll these severe storms had on people Expert structural engineers reviewed over 500 photos
in the region, the disruption of critical infrastructure from over 25 systems across five islands which were
left many communities without such basic services taken by solar professionals and system owners
as electricity for prolonged periods of time. Over the immediately after the 2017 and 2019 hurricanes. The
past decades, electricity in the Caribbean has been same structural experts from the first Solar Under Storm
primarily generated centrally by fuel oil or diesel-fired report evaluated these photographs to uncover several
engines and distributed across the island by overhead root causes of partial or full rooftop PV system failures.
lines. However, in recent years, electricity has been
supplemented by solar photovoltaics (PV) on homes,
businesses, industries, government facilities, and now, RECOMMENDATIONS
as a growing part of utility generation. In fact, over half The key output of this paper is a list of recommendations
of Caribbean electric utilities already own or operate for building more resilient rooftop solar PV systems. The
solar PV as part of their generation mix. Over 571 MW recommendations are organized into two categories:
of solar are installed across rooftops, parking canopies, 1) specifications and 2) collaboration.
and large tracts of land.5 Solar PV is the most rapidly
growing source of power for many Caribbean islands.6 1. Specifications
The following specifications list is intended as a resource
Despite the record sustained wind speeds of over for anyone who can influence project design.
180 miles per hour, many rooftop solar PV systems
in Puerto Rico, British Virgin Islands, the US Virgin • If top-down clamps are required, use clamps
Islands, The Bahamas, and Dominica survived and that hold modules individually or independently.
continued producing power. In contrast, other systems Alternately, specify through-bolting of modules.
in the region suffered major damage or complete
failure with airborne solar modules, broken equipment, • Specify bolt hardware that is vibration-resistant and
and twisted metal racking. appropriate for the environment and workforce.
Generating energy with solar PV is a cost-effective • Do not use self-tapping screws for structural
and reliable solution for power generation in the connections.
Caribbean. Incorporation of the best available
engineering, design, delivery, and operational • Specify a project QA/QC process including
practices can increase the reliability and survival rates items like bolt torqueing, ballast placement, and
from extreme wind loading. Given the variability in mechanical attachment quality.
wind speed, wind direction, wind duration, topography,
design, and construction, along with limited data, • Pitched-roof systems should only have modules
there is not an overarching statistical conclusion installed within the envelope of the roof structure (no
to explain survivorship versus failure. Instead, this overhanging modules over the roof edges) and should
SOLAR UNDER STORM PART II – 9
EXECUTIVE SUMMARY
EXHIBIT 1
Similarities of Systems
Similarities of Failed Systems Similarities of Surviving Systems
Top-down or T-clamp cascading Appropriate use/reliance on ballast
failure of module retention and mechanical attachments
Lack of vibration-resistant Sufficient structural connection
connections strength
Corner of the array overturned due Through-bolted module retention or
to incorrect design for wind four top-down clips per module
Insufficient structural connection
strength Structural calculations on record
Roof attachment connection failure Owner’s engineer with QA/QC
program
System struck by debris/impact Vibration-resistant module bolted
damage, especially from liberated connections
(dislodged) modules
Failure of the structural integrity of
the roof membrane
PV module design pressure too low
for environment
10 – SOLAR UNDER STORM PART IIEXECUTIVE SUMMARY
be limited to installation only within wind zones one 2. Collaboration
and two (see Section 4: Technical Discussion). Collaboration recommendations identify opportunities
to increase the resilience of the entire value chain and
• Ballasted-only systems are not recommended due to life cycle of solar PV projects. This requires longer-
the high risk of cascading failure modes. All systems term multi-party consideration and action intended to
should have positive mechanical attachments to the level up the current solar industry standard.
building structure that meet the minimum mechanical
attachment recommendation (see Appendix B). • Collaborate with the installer to implement and
continuously improve full QA/QC and operation and
• Require structural engineering be performed in maintenance (O&M) processes throughout the life of
accordance with ASCE 7 and site conditions, with the project.
sealed calculations for wind forces, reactions, and
attachment design. • Collaborate with professional engineers of record on
calculation best practices and intent.
• Confirm with racking vendor and project engineer
that actual site conditions comply with their base • Collaborate with racking suppliers to carry out
condition assumptions from wind-tunnel testing. full-scale and connection tests representative of
ASCE 7 3-second design wind speeds (Saffir
• Confirm with the project engineer that design best Simpson Category 5). Specifically including wind
practices are met relating to worst-case joist loading, tunnel testing review and rigidity assessment.
base velocity pressure, rigidity assessment, area
averaging, and minimum mechanical attachment • Collaborate with roofers, roofing manufacturers, and
scheme (see Appendices B and C). insurance companies to maintain roof warranty and
roof integrity.
• Require roof pre-inspections be performed to verify
that the roof conditions are acceptable and match the • Collaborate with equipment suppliers to document
assumptions in the structural design (see Appendix D). material origin and certificate of grade and coating
consistent with assumptions used in engineering
• Specify high-load (target 5,400 Pa front load rating) calculations.
PV modules, based on structural calculations; these
are currently available from a number of Tier-1 • Collaboration between installers and module
module manufacturers and tend to have more robust suppliers/distributors to ensure local availability of
frames. specified modules.
• Specify all hardware be sized based on 25 years (or
project life) of corrosion.
SOLAR UNDER STORM PART II – 111 INTRODUCTION Image courtesy of Edward Previdi, EP Energy
INTRODUCTION
Solar photovoltaic (PV) systems have proliferated During the last three years (2017–2019), the North
throughout the Caribbean and other island Atlantic region saw 11 major hurricanes (Category 3 or
communities over the past several years. Solar is higher); most notably in the Caribbean were Harvey,
now competitive with traditional fossil fuel generation Irma, Maria, and Dorian. The solar PV failures seen
and in some cases has become the primary energy from these events were well documented by the solar
source for island power systems.7 Rooftop solar has industry and serve as a continuous learning platform
also demonstrated an ability to withstand major wind on which the industry’s resilience movement stands.
events despite well documented failures. The survival and failure of ground-mounted solar
EXHIBIT 2
Guiding Principles and Process
Guiding Principles Guidelines Process
Collaborate across organizations Conduct failure analysis of sites
and integrate local experience and impacted by the hurricane seasons
expertise. from 2017 through 2019.
Address observed failure modes Engage experts responsible for
and lurking failure modes managing or analyzing historical
(ones that did not occur only because failures of solar projects.
something else failed first).
Identify and prioritize root causes
Plan for advancement of hardware, through collaborative completion of
reliability statistics, and expert a “fishbone” tool.
knowledge.
Complete a failure mode and effects
Provide performance-based analysis (FMEA) for the prioritized
recommendations where possible root causes.
to allow for innovative solutions.
Synthesize recommendations
Limit recommendations to only from the FMEA for communication
those that provide a risk-adjusted and consideration.
economic benefit.
Seek and incorporate ongoing
Ensure guidelines are executable feedback.
with currently available solutions.
Use a risk-conscious framework for
decision-making.
SOLAR UNDER STORM PART II – 13INTRODUCTION
PV systems in hurricanes was documented and well ORGANIZATION
received in the first Solar Under Storm report. Following This document is organized to present readers
the first publication, there were significant requests with each of the major analysis steps in order of
to report on the resilience of roof-mounted systems. completion. Section 2 presents the root cause
However, given the variability in wind speed, direction, identification methodology and findings, along with
roof type, roof orientation, roof pitch, solar design, recommendations for using the findings and the
and construction, one overarching conclusion cannot method. Section 3 utilizes the root causes identified in
be made to explain the diversity of outcomes from an failure mode and effects analysis (FMEA). The output
these major wind events. Instead, this report combines of this analysis includes potential mitigation actions that
photographs from immediately after the storms with are evaluated by cost and impact. Section 4 synthesizes
expert analysis to deliver actionable recommendations mitigation actions identified in the FMEA into a list
for increasing resilience among retrofit and new of recommendations for ease of communication and
construction solar PV rooftop installations. consideration by the reader.
APPROACH
Our approach to increasing the ability of PV systems to
withstand hurricane winds utilizes design-for-reliability
principles and methods.
Image courtesy of David Kaul, SALT Energy
14 – SOLAR UNDER STORM PART II2 ROOT CAUSE IDENTIFICATION Hurricane Dorian September 2019. Image courtesy of NOAA
ROOT CAUSE IDENTIFICATION
The rise in hurricane intensity in the region in manufacturing, procurement, delivery, installation, and
conjunction with the increased installed base of operations of a rooftop solar power plant, along with
solar PV has provided an initial body of evidence for the operational use case. The most urgent causes
developing resilience guidelines for future projects. of failure are in bold text. The current fishbone draft
However, development of hurricane resilience is limited by the data set, authors’ expertise, and
guidelines based on observed failure modes alone current technology; consequently, this analysis should
has limitations. The observed failure modes may have be updated to incorporate new data, expertise, and
served as a “mechanical fuse” relieving forces from the technology. Future solar PV project teams are invited
system. If future systems only address the observed to utilize Exhibit 3 (and add additional categories as
failure modes, forces may precipitate additional failure necessary) as a facilitation tool to explore project-
modes. To address both observed and potential specific opportunities to eliminate causes of failure in
failure modes, we took a classic reliability engineering response to extreme wind or other hazards.
approach to design for reliability. Exhibit 3 illustrates a
common reliability tool for systematic cause and effect
identification called a fishbone diagram. The diagram
shows the supply chain responsible for design,
Image courtesy of Carlos Quiñones, CJQ Engineering
16 – SOLAR UNDER STORM PART IIROOT CAUSE IDENTIFICATION
EXHIBIT 3
Fishbone diagram
Building
Roof Features &
Equipment
Equipment Stakeholders
Roof Properties
Module
Warranties
Construction
Roof Age
Material
Load Rating Management O&M Provider
Type
Materials
Module Hardware Documentation Asset Owner
Metal Grade Design Life Training Developer Building Properties
Structural
Capacity
Coatings Reliability Testing Means & Methods Local Inspectors
Height
Speed
Pitch
Quality
Certifications Warranty Management Installer / EPC
HURRICANE FAILURE
Equipment
Corrosivity Roof Wind Zones Code Enforcement Transaction System Grounding
Worst Case Component Construction Electrical
Topography Joist Loading Certifications Agreement Connectors
Layout Electrical Wire
Wind Obstructions Service Agreement Management
TUV
Turbulence
IEC
Wind Tunnel Test
UL
Procurement
Exposure
Direction
Report Process
Speed
Electrical
Ballast
Placement (4) System Certifications
Member & Business Model
Environment
ASCE 7
Connection Sizing
SEAOC
AISC
Mechanical Attachments
Effective Wind Area
Ballast Placement
Roof Wind Zone
IBC
Codes and
Standards
System Design
SOLAR UNDER STORM PART II – 173 FAILURE MODE AND
EFFECTS ANALYSIS
Image courtesy of of Carlos Quiñones, CJQ EngineeringFAILURE MODE AND EFFECTS ANALYSIS
Improving the ability of PV systems to withstand 1. Top-down clip failure (all projects)
hurricane winds requires not only identification of failure 2. Debris/impact failure (all projects)
modes but also a cost-effective mitigation action. The 3. Corner overturn failure (ballasted/mechanically
failure mode and effects analysis (FMEA) framework attached hybrid flat roof)
was utilized to identify practical mitigation actions. This 4. Racking connection failure (all projects)
assessment is a culmination of two markers: expected
upfront material cost to implement the mitigation actions Debris/impact failure was largely a secondary failure
and the impact this mitigation will have on total cost of mode caused by clip failures. Eighteen projects
ownership (TCO). TCO reductions are driven through (70%) experienced debris/impact failures while only
a reduction in total economic two of those projects (8%)
damage or a reduction in the Out of this data set, “top- experienced impacts from
frequency of individual failure objects other than liberated
modes. Reduction in total down clip failure” was modules. Mitigating the
economic damage directly observed on 21 out of cascading failure mode by
improves long-term asset
22 projects that utilized solving the “top-down clip
values by minimizing material failure” largely eliminates this
replacements. Reduction in top-down clips (96% of failure mode as well.
frequency of failures improves applicable projects).
PV plant up-time by minimizing Corner overturning was
the time spent fixing minor only present on ballasted/
issues, especially cascading issues that could lead to mechanically attached hybrid flat-roof projects. However,
major, expensive failures if not immediately addressed. it was present on three of the four projects within the
data set (75%). Due to the limited number of ballasted/
The synthesis of the FMEA presented below is mechanically attached hybrid projects (four), it is difficult
designed to teach a user the current practices and to extrapolate these failure modes past the observed
associated limitations of the most relevant failure modes portfolio. However, the root causes in these photos are
and to provide a cost-effective mitigation action. The evident and are supported by the FMEA activities.
table is organized by subsystems and assemblies.
Racking connection failures speak to the compromised
In addition to the FMEA work, we performed a structural integrity of the racking system itself. It
statistical analysis on a limited data set (26 total occurred on 6 out of 26 projects (23%), including the
rooftop projects) of available failure images. Projects 3 projects which experienced corner overturning
included sloped pitched-roof structures (11 projects), failures. Only on a single occasion was a racking
elevated rail system flat-roof structures (11 projects), connection failure deemed to be a primary failure
and ballasted/mechanically attached hybrid flat-roof mode. Mitigating the cascading failure mode by
structures (4 projects). Four major failure modes solving the “corner overturning failure” largely
presented themselves within this data set. They are eliminates this failure mode as well.
listed in order of decreasing occurrence:
SOLAR UNDER STORM PART II – 19FAILURE MODE AND EFFECTS ANALYSIS
EXHIBIT 4
Cost/Impact Key
Total Cost of Ownership
Cost ($/watt)*
(TCO) Impact
LowFAILURE MODE AND EFFECTS ANALYSIS
EXHIBIT 5
Failure Mode and Effects Analysis—Buildings
Failure Modes Current Practice Limitations Potential Mitigation Cost/TCO
Incorrectly calculated ASCE 7-16 Relies on project Third party review of the ASCE Low/High
velocity pressure engineer to properly 7-16 velocity pressure calculation.
capture project-
specific factors.
Roof structural Project engineer Requires correct Require “worst-case joist loading” Low/Med
member failure checks dead loading identification of be checked in addition to “array
against available governing load cases area averaging” (see Appendix C).
capacity. and roof capacity.
Wind acceleration in • Racking vendor Racking vendor • Adopt minimum mechanical Med/High
specific wind zones specifies wind analysis can be a attachment specification (see
loading. “black box” to project Appendix B).
• ASCE 7-16 engineer. • Must be considered in wind
tunnel testing for flat-roof
projects.
• Require installation of modules
only in Zones 1 and 2 for
pitched-roof systems (see
Section 4: Technical Discussion).
All Images courtesy of Carlos Quiñones, CJQ Engineering
SOLAR UNDER STORM PART II – 21FAILURE MODE AND EFFECTS ANALYSIS
EXHIBIT 6
Failure Mode and Effects Analysis—PV Racking
Failure Modes Current Practice Limitations Potential Mitigation Cost/TCO
Cascading failure of • Module top-down Shared (middle) • Use top-down clips that do not Low/High
top-down clips clips are designed module top-down clips retain more than one module
to retain groups of lose capacity with per clip to avoid cascading
modules with end loss of one module failures.
and mid (shared) and allow liberation of • Alternately, specify module
clamps. adjacent module. frames to be through-bolted in
• Caribbean regional accordance with manufacturing
solution of a three- specification for the design wind
rail system has been speed.
popular.
Fastener self- Racking vendor Wind vibrations can • Require vibration-resistant Low/High
loosening selects the hardware loosen hardware over fasteners.
in the design. time. • Ensure proper QA/QC during
installation.
• Verify tight fasteners during
annual O&M activities.
Roof to racking Racking vendor • Incorrect loading • Verify racking vendor meets Low/High
mechanical performs structural assumed (effective recommended analysis
attachment failure calculations wind area) invalidates and minimum mechanical
calculations. attachment scheme
• Improper installation. (see Appendix B).
• Proper installation QA/QC is
critical.
Self-tapping screw Racking vendor • Subject to improper • Avoid self-tapping screws for Low/Med
corrosion and failure selects the hardware installation. structural loading in the design.
in the design. • Corrosion happens
over time.
All Images courtesy of Carlos Quiñones, CJQ Engineering
22 – SOLAR UNDER STORM PART IIFAILURE MODE AND EFFECTS ANALYSIS
EXHIBIT 6 (CONTINUED)
Failure Mode and Effects Analysis—PV Racking
Failure Modes Current Practice Limitations Potential Mitigation Cost/TCO
Corner overturning Racking vendor Incorrect loading • Verify racking vendor meets Med/High
failures performs structural assumed (effective recommended analysis
calculations wind area) invalidates and minimum mechanical
calculations. attachment scheme (see
Appendix B).
Racking uplift or Racking vendor Incorrect loading • Verify racking vendor meets Med/High
sliding failures performs structural assumed (effective recommended analysis
calculations wind area) invalidates and minimum mechanical
calculations. attachment scheme (see
Appendix B).
Incorrectly calculated Racking vendor Racking vendor • Adopt minimum mechanical Med/High
module-specific wind specifies wind loading. analysis can be a attachment specification (see
loads “black box” to project Appendix B).
engineer.
Dynamic excitation— Racking vendor or High difficulty in • Adopt minimum mechanical Med/Med
“Walking” of the project engineer designing a purely attachment specification (see
ballasted rack performs calculations. ballasted racking Appendix B).
system around • Project engineer should review
dynamic wind structural elements for thermal
considerations. expansion considerations and
seismic loading.
Wind deflector Specific to racking • Vulnerable to • Install positively retained Low/High
liberation design. installation errors. wind deflector with vibration-
• Vulnerable to impact resistant solution.
damage. • Proper QA/QC at installation
• Vulnerable to • Verification during annual O&M
improper design. activities
All Images courtesy of Carlos Quiñones, CJQ Engineering
SOLAR UNDER STORM PART II – 23FAILURE MODE AND EFFECTS ANALYSIS
EXHIBIT 7
Failure Mode and Effects Analysis—Electrical
Failure Modes Current Practice Limitations Potential Mitigation Cost/TCO
Wire pull out or Specification Terminal Specify QA/QC procedure and Low/Low
terminal damage for each torque values documentation for terminal
electrical unchecked in torques.
Component (e.g., UL, field.
NEC, TUV, etc.).
Wire sheath chafing NEC or IEC conductor Wires sag and subject • Evaluate racking structures for Low/Low
(ground fault) management and to gyration based on inclusion of wire management
support specifications. field installation. solutions.
• Proper QC of field electrical
work.
• Specify wire management
practices, including support
schedule and sag tolerance.
• Specify stainless-steel or heavily
galvanized wire clips or PVC-
coated stainless-steel cable
clamps instead of plastic zip ties.
Rain intrusion NEC - NEMA Hurricane force wind • Verify water sealing method Low/Low
specification can drive rain. effective at the project design
wind speed.
Wind load on electrical ASCE/NEC/UL Codes Combiner boxes, • Have the project engineer Low/Med
components inverters, and other analyze electrical components
equipment are and their structural mounting
exposed to wind loads to resist applicable project
but rarely analyzed or environmental loads.
properly secured.
All Images courtesy of Carlos Quiñones, CJQ Engineering
24 – SOLAR UNDER STORM PART IIFAILURE MODE AND EFFECTS ANALYSIS
EXHIBIT 8
Failure Mode and Effects Analysis—PV Modules
Failure Modes Current Practice Limitations Potential Mitigation Cost/TCO
Frame bolt hole failure UL1703 certification of Module back-side • Specify engineer calculations Low/Low
module testing. (uplift force) rating may for module connection
not be adequate for hardware, including frame
local loads. where used.
• Collaborate with module
manufacturers to improve
supply chain.
• Engineer of record for the
project should request
and approve engineering
connection calculations.
Laminate impact UL 1703 hail impact Hurricane debris can • Specify that site prep and Low/Med
damage tests and ASCE wind be large compared to clean-up shall include removal
prone debris. hail. or securement of all foreign
objects (debris).
• Execute proper failure mode
mitigations for module
liberation, especially top-down
clips and vibration-resistant
hardware.
All Images courtesy of Carlos Quiñones, CJQ Engineering
SOLAR UNDER STORM PART II – 254 TECHNICAL DISCUSSION Image courtesy of Pura Energía
TECHNICAL DISCUSSION
THREE-RAIL SYSTEMS
Pitched-roof system designers in the region have loss of any corner clip would result in wind pressures
often utilized a “three-rail system” as a viable solution overstressing remaining clips. The cascading failure
against module liberation on pitched-roof systems. mode mitigation recommendation of moving away from
This is driven by a requirement for 5400 Pa front load shared top-down clips is still a valid recommendation as
rated modules. Module manufacturers that offer such a additional clips don’t solve the root problem.
warrantied rating do so by often requiring six module
fasteners instead of the existing requirement of four. The six-fastener requirement of module manufacturers
The additional two fasteners require an additional rail to obtain a 5,400 Pa rating is an externalization of cost
for mounting. onto the projects of the Caribbean region. Material
cost increases incurred by simply supplying a heavier
Anecdotally, designers have cited benefits of a three- module frame would certainly be lower than a 50%
rail system being used to obtain a reduction in vibratory increase in project racking material and mechanical
forces in the module leading to less chance of module attachments. Module frame manufacturing is done via
liberation as well as providing added structural strength aluminum extrusion, which is a very materially efficient
toward resisting extreme wind forces. Top-down clips in process, especially at high volumes.
this application have also performed better at retaining
modules during hurricanes. Thus, a three-rail system represents a significant cost
increase for projects in the region (50% increase in
Although three-rail systems do provide these benefits, racking structure and roof attachments) while aiming
they solve module liberation issues that are rooted to solve problems that would be better solved with
in other failure modes. Addressing the root cause of appropriate fastener selection. Some of this is driven
the failure is important. For example, hardware that by module manufacturers requiring higher sales
vibrates loose should be replaced by vibration-resistant volumes for high-wind zone projects to invest in a
hardware and not simply additional hardware. Top-down dedicated 5,400 Pa rated frame. But some of it is
clips perform better because the root cause of their rooted in racking manufacturers selling additional
failure is vibratory. It is more effective to solve the root racking rather than selecting new hardware. Owners
cause in a cost-efficient way than to invest in material to and project engineers should be selective in this
reduce the vibration in the system. In fact, the structural regard, scrutinizing hardware selection and pushing
calculations on a six-fastener system show that the for the most cost-efficient solutions.
Three-rail system on St John, St. John, US Virgin Islands. Image courtesy of Caribbean Solar Company
SOLAR UNDER STORM PART II – 27TECHNICAL DISCUSSION
JOINT LOOSENING / TOP DOWN CLIP FAILURE ANALYSIS
Most structural failures of PV systems originate at load tension is reduced. A third mode of joint loosening
the connections. As highlighted in several examples is when transverse vibration causes joint loosening
in the FMEA analysis above, loosening of joints is by rotation of a nut or other fastener. This mode has
often a contributing factor to connection failure. The gained the most attention, as it is easiest to observe by
probability of joint loosening can be reduced by visual inspection. Joint loosening caused by transverse
proper selection, specification, and installation of vibration can be simulated with the Junker vibration test.
fasteners, to avoid loss of pre-load tension (clamping Some forms of lock nuts or lock washers that have been
force). There are a few common failure modes to be trusted for years in multiple industries can be shown
considered. When selecting and specifying fasteners, to spin off during the course of a Junker transverse
bear in mind three primary concerns. Select fasteners vibration test, while a certain type of two-piece stepped
that will maintain their pre-load tension; the fasteners washer performs well in the test. It is important to
must have adequate corrosion protection to survive recognize that transverse vibration is only one mode
the life of the PV system; and they must be compatible of concern, and each installer should perform their
with electrical bonding and grounding concerns. As own due diligence on final solutions for selection and
there are multiple criteria for selection of fasteners, specification of fasteners in high-wind installations.
we recommend that no substitutions be allowed in
the field once fasteners are specified without express During installation, if a telltale mark is added to
written consent for an alternative specification from fasteners after initial torque setting, then O&M
the structural engineer of record. personnel can visually observe whether the marks
continue to align or if the joint has experienced rotation.
There are at least three modes that can contribute to
loss of pre-load tension and associated joint loosening.
The first mode of concern is “pre-load scatter,” which
refers to an unintentional variation in initial torque of
fasteners. While several references will recommend that
all fasteners be individually set with a torque wrench,
it is more likely that large quantities of fasteners will be
installed with a calibrated, torque-controlled driver. To
minimize pre-load scatter, a quality assurance program
could include a data-logging torque-controlled driver,
such that a record of initial torque is created and
maintained with operations and maintenance (O&M)
personnel. A second mode of concern is embedment
of lock washers in the base material. For example, a
common, low-cost star washer is likely a poor choice for
bonding and grounding, as repeated vibration cycles
can cause additional embedment in the base metal,
causing a loss of pre-load. In this case, connections can
experience joint loosening without any turn of a nut.
Joint loosening caused by additional embedment after Image courtesy of Christopher Burgess, Rocky
initial torque can lead to further loosening after pre- Mountain Institute
28 – SOLAR UNDER STORM PART IITECHNICAL DISCUSSION
MODULES OVERHANGING THE ROOF EDGE
Installing modules over the edges of roofs is another Fire codes typically require rooftop setbacks and
regional practice often encountered. This maximizes access pathways for rooftop operations for fire
the array size for the customer, offsetting greater fighters. For commercial buildings with low-slope
consumption and lowering the total cost of the project roofs, these fire setbacks are often 4 feet or 6 feet
(on a per-watt basis). However, these modules are almost from roof edges, depending on the size of the
universally missing post-hurricane. The mitigation isn’t building. For high-slope roofs, fire setbacks and
just rooted in common sense, but also in building code. access pathways are often 3 feet from ridges and
gable ends but could be as little as 18 inches or not
For rooftop PV systems, the proximity of the PV panels required at all, depending on configuration of the
to roof edges is of primary concern. As wind flows roof and requirements of a local fire service. For
toward a building and is obstructed by the building, PV systems installed parallel to a roof, ASCE 7-16
it must travel around the wall corners and up over requires a setback from all roof edges that is at least
the roof. As wind travels over the roof edge, it can twice the height of the PV panels above the roof. For
“detach” from the roof, resulting in a large negative example, if a PV system is installed 5 inches above
(uplift) pressure near the roof edge. For large buildings and parallel with a high-slope roof, it must be set
with low-slope roofs, the wind pattern reattaches back at least 10 inches from all roof edges. To reduce
to the roof at some point farther downwind. As the probability of failure owing to unanticipated wind
highest uplift wind pressures occur near roof edges, uplift pressures, PV system layout should avoid any
the design layout should consider some setback of PV overhangs above a roof ridge and should be set back
modules from roof edges. from all roof edges.
Modules Overhanging the Roof Edge. Image courtesy of Marc Lopata, Solar Island Energy
SOLAR UNDER STORM PART II – 29TECHNICAL DISCUSSION
EXHIBIT 9
Wind Loads—Components and Cladding
Exhibit 9 shows a relatively lower pressure region ASCE) due in part to Hurricanes Irma and Maria. Any
(green - Zone 1), medium pressure regions (yellow - modules within the envelope of Zones 1 and 2 should
Zone 2), and high pressure regions (red - Zone 3). have the appropriate code-provided wind pressures
applied and professional engineers should check that
PV array designs should stay within the envelope of the structural stability of the system and its connection
the roof boundaries, but they should also universally to the roof is adequate. Modules within Zone 3 would
stay away from installation in Zone 3. The wind receive such catastrophically high wind pressures
pressure in this zone is significantly higher (~50%) that the cost of installing them appropriately for these
than in Zone 2. This is especially true in design loads would be universally untenable.
wind speeds of 120 mph and above, which many
municipalities of the Caribbean are rated at (or soon
will be upon the latest release of wind maps from
30 – SOLAR UNDER STORM PART IITECHNICAL DISCUSSION
TOPOGRAPHIC “SPEED-UP” EFFECTS FOR DETERMINATION OF DESIGN WIND SPEED
Determination of the correct design wind speed is Rico in ASCE 7-16 still show the “old” maps with only
fundamental to calculating design wind pressures a few contours. The Applied Technology Council
using ASCE 7-16. When wind speed increases with (ATC) tool for Puerto Rico is now online and is critically
changes in topography, it can have a strong influence important to use. (Do not use the maps in ASCE 7-16
on design wind pressures. Forensic studies in Puerto for Puerto Rico.) It is important to note that the ATC
Rico after Hurricanes Irma and Maria determined online tool for Puerto Rico (only) includes wind speed-
some failures were partially attributed to absence of up effects, so the topographic factor is already built
consideration of topographic effects in determination in. For determination of design wind pressures other
of design wind pressures. than for Puerto Rico, it is important to calculate the
topographic factor. (This method is being revised
After Hurricanes Irma and Maria, FEMA funded a during the development of future ASCE 7-22, so
wind-speed study for the entire island of Puerto Rico, sophisticated readers might want to look at the in-
with a goal of providing a more-accurate wind speed progress changes.) At the time of writing this paper, a
map and an online tool for determination of design similar wind speed study is being conducted for the
wind speed, similar to the effort for the islands of US Virgin Islands.
Hawaii years ago. The wind speed maps for Puerto
SOLAR UNDER STORM PART II – 315 CONCLUSION Image courtesy of FortisTCI, Turks and Caicos
CONCLUSION
Generating energy with solar PV is a cost-effective its ability to be omniscient of all failure modes and all
and reliable solution for power generation in the corrective actions and cannot guarantee the efficacy
Caribbean. Incorporation of the best available of any recommended action. However, it provides
engineering, design, delivery, and operational a set of best practices regarding specifications of
practices can increase the reliability and survival rates equipment and procedures along with a framework for
from extreme wind loading. This paper is limited in continued collaboration.
EXHIBIT 10
Similarities of Systems
Similarities of Failed Systems Similarities of Surviving Systems
Top-down or T-clamp cascading Appropriate use/reliance on ballast
failure of module retention and mechanical attachments
Lack of vibration-resistant Sufficient structural connection
connections strength
Corner of the array overturned due Through-bolted module retention or
to incorrect design for wind four top-down clips per module
Insufficient structural connection
strength Structural calculations on record
Roof attachment connection failure Owner’s engineer with QA/QC
program
System struck by debris/impact Vibration-resistant module bolted
damage, especially from liberated connections
(dislodged) modules
Failure of the structural integrity of
the roof membrane
PV module design pressure too low
for environment
SOLAR UNDER STORM PART II – 33CONCLUSION
RECOMMENDATIONS
The key output of this paper is a list of recommendations • Confirm with racking vendor and project engineer
for building more resilient rooftop solar PV systems. The that actual site conditions comply with their base
recommendations are organized into two categories: condition assumptions from wind-tunnel testing.
1) specifications and 2) collaboration.
• Confirm with the project engineer that design best
practices are met relating to worst-case joist loading,
1. Specifications base velocity pressure, rigidity assessment, area
The following specifications list is intended as a resource averaging, and minimum mechanical attachment
for anyone who can influence project design. scheme (see Appendices B and C).
• If top-down clamps are required, use clamps • Require roof pre-inspections be performed to verify
that hold modules individually or independently. that the roof conditions are acceptable and match the
Alternately, specify through-bolting of modules. assumptions in the structural design (see Appendix D).
• Specify bolt hardware that is vibration-resistant and • Specify high-load (target 5,400 Pa front load rating)
appropriate for the environment and workforce. PV modules, based on structural calculations; these
are currently available from a number of Tier-1
• Do not use self-tapping screws for structural module manufacturers and tend to have more robust
connections. frames.
• Specify a project QA/QC process including • Specify all hardware be sized based on 25 years (or
items like bolt torqueing, ballast placement, and project life) of corrosion.
mechanical attachment quality.
• Pitched-roof systems should only have modules
installed within the envelope of the roof structure (no
overhanging modules over the roof edges) and should
be limited to installation only within wind zones one
and two (see Section 4: Technical Discussion).
• Ballasted-only systems are not recommended due to
the high risk of cascading failure modes. All systems
should have positive mechanical attachments to the
building structure that meet the minimum mechanical
attachment recommendation (see Appendix B).
• Require structural engineering be performed in
accordance with ASCE 7 and site conditions, with
sealed calculations for wind forces, reactions, and
attachment design.
34 – SOLAR UNDER STORM PART IICONCLUSION
2. Collaboration
Collaboration recommendations identify opportunities
to increase the resilience of the entire value chain and
life cycle of solar PV projects. This requires longer-
term multi-party consideration and action intended to
level up the current solar industry standard.
• Collaborate with the installer to implement and
continuously improve full QA/QC and operation and
maintenance (O&M) processes throughout the life of
the project.
• Collaborate with professional engineers of record on
calculation best practices and intent.
• Collaborate with racking suppliers to carry out
full-scale and connection tests representative of
ASCE 7 3-second design wind speeds (Saffir
Simpson Category 5). Specifically including wind
tunnel testing review and rigidity assessment.
• Collaborate with roofers, roofing manufacturers, and
insurance companies to maintain roof warranty and
roof integrity.
• Collaborate with equipment suppliers to document
material origin and certificate of grade and coating
consistent with assumptions used in engineering
calculations.
• Collaboration between installers and module Image courtesy of Fidel Neverson, Energy Solutions, Inc.
suppliers/distributors to ensure local availability of
specified modules.
SOLAR UNDER STORM PART II – 35ENERGY STORAGE SYSTEMS FOR RESILIENCE
While this paper is focused solely on solar PV systems, this additional resilience can ensure continued critical
it is worth adding that PV systems combined with a services to the community such as communications,
battery storage system can continue to deliver power water treatment and pumping, medical operations, and
to a home, business, or critical facility, even during a refrigeration for food and medicine storage.
grid outage. Most grid-connected PV systems without
battery storage will shut down when a grid outage is By pairing batteries with a resilient solar PV system,
detected, to avoid back-feed to the grid and to ensure facilities can count on uninterrupted power even after
safety of the system and utility personnel. A PV system the most severe storms. Additional discussion on the
with a multi-mode inverter, transfer switch, battery many benefits of solar coupled with battery energy
storage system, and other appropriate components storage can be found on RMI’s blog post “Critical
can be disconnected (“islanded”) from the grid during Facilities: Where Government and Utility Services
a power outage. During extended power outages, Redefine Resilience.”8
Image courtesy of The Solar Foundation
36 – SOLAR UNDER STORM PART IIRECOMMENDED REFERENCES Image courtesy of Rocky Mountain Institute
RECOMMENDED REFERENCES
FEMA Advisory: Rooftop Solar
USVI - RA 5 - Rooftop Solar Panel Attachment: Design,
Installation, and Maintenance
https://www.fema.gov/media-library/assets/
documents/158123
FEMA P-2021 | Mitigation Assessment Team
Report: Hurricanes Irma and Maria in the U.S.
Virgin Islands
https://www.fema.gov/media-library/assets/
documents/170486#
FEMA P-2054 | Mitigation Assessment Team
Compendium Report
https://www.fema.gov/media-library/assets/
documents/184600
ATC Wind Hazard Tool
https://hazards.atcouncil.org/#/wind
Minimum Design Loads and Associated Criteria
for Buildings and Other Structures (ASCE/SEI 7-16)
https://www.asce.org/asce-7/
Solar Photovoltaic Systems in Hurricanes and
Other Severe Weather, US Department of Energy
https://www.energy.gov/sites/prod/files/2018/08/
f55/pv_severe_weather.pdf
Image courtesy of Caribbean Solar Company
38 – SOLAR UNDER STORM PART IIYou can also read
